基于句子语义挖掘的上市公司风险事件知识聚合模型  

Knowledge Aggregation Model of Listed Companies′Risk Events Based on Sentence Semantic Mining

作　　者：谭明亮 游强华 杨达森 周禾深 唐晓波[3] TAN Mingliang;YOU Qianghua;YANG Dasen;ZHOU Heshen;TANG Xiaobo(School of Management,North Sichuan Medical College,Nanchong 637100,China;不详)

机构地区：[1]川北医学院管理学院,四川南充637100 [2]中国人民大学信息资源管理学院,北京100872 [3]武汉大学信息管理学院,湖北武汉430072

出　　处：《武汉理工大学学报（信息与管理工程版）》2022年第1期158-165,共8页Journal of Wuhan University of Technology：Information & Management Engineering

基　　金：川北医学院博士启动金项目(CBY21-QD37).

摘　　要：大数据环境下,对上市公司风险事件相关的多源异构文本数据资源进行有效的挖掘和聚合,使其收敛于高质量的金融知识服务,对于提升证券投资分析、资产风险管理、证券市场监管等金融管理决策的效果和效率有着重要的意义。文章以文本挖掘、知识组织和知识服务等领域的理论、方法和技术为基础,构建了基于句子语义挖掘的上市公司风险事件知识聚合模型,该模型由数据采集与数据预处理模块、知识获取与知识组织模块、知识利用与知识服务模块所构成。文章通过构建并利用中文金融情感词典来提取上市公司风险事件相关的临时公告和新闻报道,应用LDA模型和BERT模型来挖掘句子的语义特征,并利用凝聚层次聚类算法来实现句子的聚类,然后选择重要度高的句子生成上市公司风险事件的主题报告。Under the big data environment,the effective mining and aggregation of multi-source heterogeneous text data resources related to the listed companies′risk events makes it converge to high-quality financial knowledge services,which is of great significance for improving the effect and efficiency of financial management decisions such as securities investment analysis,asset risk management and securities market supervision.Based on the theories,methods and technologies in the fields of text mining,knowledge organization and knowledge service,this paper constructs a knowledge aggregation model of listed companies′risk events based on sentence semantic mining.The model consists of data collection and data preprocessing module,knowledge acquisition and knowledge organization module,knowledge utilization and knowledge service module.This paper constructs and utilizes a Chinese financial sentiment dictionary to extract the temporary announcements and news related to the listed companies′risk events,and uses the LDA model and the BERT model to mine the semantic features of sentences.This study utilizes the agglomerative hierarchical clustering algorithm to cluster the sentences,and then selects the sentences with high importance to generate the topic report of the listed companies′risk events.

关 键 词：句子语义挖掘 金融情感词典 上市公司风险事件 知识聚合 知识组织 

分 类 号：G356[文化科学—情报学]